Jump to content

Recommended Posts

PhysX Challenge

 

This thread originated when I was trying to save my addons when they all had stopped working after the TacOps update. Now in hindsight I want to share some examples and ideas about physX for aircraft, especially for tail drag setup.

They are intended as a reference for anyone who may be interested. Feel free to use the code, change it - and make suggestions or share insights in this thread.

 

The TacOps update of November 2017 changed the ground handling of airplaneX simulation. This probably happened accidently while BI was trying to improve tank movement (I do believe it is this change in the patch notes >Tweaked: Enabled PhysX wheel sweep collisions<). Since then aircraft set up in the traditional way as in Arma 2 and with airplaneX enabled may start to jump and blow up on the runway. This usually happens at ground speeds exceeding 50 km/h (even on VRs flat surfaces).

 

In some aspects that old system was superior to the new physX wheel setup. You could damage hide parts of the gear so you had to land on two wheels, actually simulate belly landings etc.

Some of my secret weapons planes do work this way still (the IL-2  for instance looses a wheel after some damage and can belly crash land. I have no idea why it still works and never managed to reproduce that setup).


here are some of my personal observations concerning airplaneX and physX wheels -

Not pretending to understand what is really going on. Just some observations after doing this to a few dozen aircraft:

Spoiler

For modern tricycle gear setup (http://en.wikipedia.org/wiki/Tricycle_landing_gear)

 

Resolution LOD (Tricycle)

wheels and animations here are purely visual and irrelevant to the behaviour of the plane. Fiddle with this only once your plane moves correctly. - Add at a later stage or use as reference.

If you set up the animation sources and the sources in model.cfg for the respective dampers this can help you see what is going on. Keep in mind that the virtual physx damper only has one vector of movement.

Dampers with hinges or complex dampers with many parts will have to be made in a way so they sort of match this. Tricky.

- no need to have actual wheels or gear in here really. Just for the show.

 

GEO LOD (Tricycle)

The only use I can observe with making geo lod wheels is with placing the aircraft in the editor and collision with other objects. Some of my aircraft don't have any and work just fine.

Make them smaller than the actual wheel so they do not interfere (which they don't seem to). There is no need for assigning mass to them. You do that in the physX config anyway.

Set the wheels to have the same name as in res LOD, name them after the dampers or the non moving part of the gear, or have no name attached at all. All have slightly different effects.

- Something to ladd later to prevent this from interfering with your basic setup.

 

You can use a few mass points copy/pasted from the example p3ds and have the actual model just being shape without any weight attached to it.

Remember to always update with Structure-Topology-Find Components or it wont have any effect.

If you intend to use BIs buzzard cannon, have the total weight set above 5tons or you plane will fly backwards from the recoil.

Apart from that total mass is only used for collision with other planes or trucks on the ground.

It does reduce maneouverability but you could have a 1ton plane behave like a 100ton plane by reducing aileron and elevator sensitivity.

- So I use the same weight for all of my aircraft. Otherwise you will have to redo damper settings over and over again, if you happen to make more than 1 aircraft.

 

GEO PHYSX LOD (Tricycle)

Copy of geo lod sans wheels. Seems to be irrelevant for collisions between vehicles. Quite probably only needed for collision detection with thrown physX objects like grenades.

- I would suggest not using till everything else works.

 

LANDCONTACT LOD (Tricycle)

I would suggest trying without this first till your plane works. then you can still add more layers of confusion and problems by adding this later.

Seems to help with placing in editor (tail drag may have its tail floating till the engine is turned on if missing this, not a big deal). 

You can name the points after the dampers, the gear or have no name attached at all. All have slightly different effects

-Can sometimes prevent rolling away for taildraggers. Some planes work without this LOD in the p3d at all.

 

MEMORY LOD (Tricycle)

the wheel_center and wheel_rim points do not animate! They are either there (gear extended) or hidden (gear retracted).

All damping happens in relation to these points by using the damper vector and droop and compression values. that is all.

Your nice visual dampers have no effect on this. That makes complicated dampers with hinges instead of pistons a little of a challenge to get right.

Trial and error here to make it match up

- This will also enable you to use 3d models with pre folded gear though. Just place the wheel points where the gear would be if extended and reverse your visible animations to make it match up.

 

 

For tail draggers (https://en.wikipedia.org/wiki/Conventional_landing_gear):

 

make the two front wheels the steering wheels (thanks to Reyhard for this idea). Tail wheel should not be a steering wheel!

Have MOI or drag on the tail wheel. Bumper balance between front and tail is important too.

 

GEO LOD (tail drag)

Geo Lod wheels are not really necessary. My setups sometimes have geometry wheels for better placement in the Editor. That is the only effect I am certain of. Maybe add once everything else works.

If your plane rolls away on its own - move GoC against the direction of the movement and try again (name the selection after wheels, dampers or nothing at all).

Another thing that can sometimes help the tail draggers is to move the landcontact point of the tail wheel up into the wheel position. Sometimes a bit to the front of it. That seems to act as a kind of brake.

 

A tail dragger takes off at about 20-30 km/h less than a tricycle if everything is equal otherwise. The game engine obviously interprets it as some kind of high AoA.

Sometimes if your plane can not take off, it makes sense to tilt the whole setup - playing around with envelope is probably easier.

 

change of mass:

With commands and adding cargo we can affect the weight of a plane. Tricycles will roll away when GoC changes due to the additional cargo weight.

Mass change will change control sensitivity in flight (more mass = very much more agile - yes, that is not a typo!

So if anyone wants to simulate the plane becoming easier to handle after dropping something, raise the mass.

 

 

 

Final Thoughts:

For a standard tricycle gear plane all of this is not entirely bad. It does not add anything much except a lot of work and maybe some nicer damper animations. You lose the option to damage the gear functionally or to belly land though.

 

 

For a taildragger you could try the old Arma2 setup without this and use the old system first and see if you can make it work.

For slow planes (Biplanes/Ultralights etc) just stick to the old way and forget about class Wheels.

The bumping/exploding effect usually happens above 50km/h so may not be relevant to you! (examples for this are my An-2, Ultralight, and every plane in my Flying Circus and Flying Machines addons)

 

If you start a new plane with physX my suggestion is to start from bottom up. Forget about landcontact lod, complicated animations, geo lod wheels or damper setup.

Make the absolute minimum work first, then move upwards. The basics are the wheel_center and wheel_rim memory points and your class Wheels. Forget about dampers. set the droop and compression to 0. Completely remove the gear from res lod and geo lod so you don't start believing it actually does anything. Then add layers of complexity as needed. That way you won't spend days adjusting something that isn't even the cause of an issue you may encounter.

 

 

 

 

 

DOWNLOAD example aircraft

Ca-12 Boomerang Tail Drag setup

 

There is an edit layer with all relevant parts all in one. Adjust the wheel positions, - then use the named selections to select the parts to copy/paste into GEO,LC and Memory LOD.

I will re-add a tricycle example some other time

 

 

 

Short guide for implementing this in your plane:

Spoiler

Make driveoncomponents an empty array.

Add the class wheels to your config.

Add wheel_center and wheel_rim to your memory lod. where the wheels would be with the gear down and touching the ground. remove geo lod wheels for time being, maybe even landcontact lod.

Add custom animation sources for the dampers (class animationsources). I personally do not use the new animation source for wheel rotation. The old one works fine.

Link those to your animation in the model.cfg.

Make sure the totalweight matches my examples if using that code as base. otherwise the plane will jump all over the place or roll off.

 

 

Thanks to TeTeT and Reyhard for some great suggestions!

  • Like 3

Share this post


Link to post
Share on other sites

Looks interesting. May I ask what the actual challenge is?

 

I'm guessing its getting a working setup that doesnt get thrown across the map? 

Share this post


Link to post
Share on other sites

Yes, the challenge is to make a plane that does not explode when trying to taxi or take off/land at speeds exceeding 50 km/h.

For a tricycle that can be done, but for a tail dragger it's a real nightmare.

 

We have a first entry too. TeTeTe3 was so kind to make a pretty decent setup for both. The tail dragger is a little bumpy still but both are far better than anything I could do.

I will provide a link sometime I get around to it. Right now physX got me so fed up again I need a break from it.

 

 

Share this post


Link to post
Share on other sites

Have you tried to configure two front wheels as rotating one? It will be fake, indeed, but I guess nobody will notice that without PhysX diagnostic.

  • Like 1

Share this post


Link to post
Share on other sites

Hi Reyhard, thx for dropping by!

I did something similar, making a fake invisible front wheel. Thinking about it now that was a really silly idea.

All kinds of issues arose from not having the thing leveled out.The taildragger becomes tilted upon spawn naturally. The tail dropping to the ground. That seemed to cause all kinds of weird behaviour.

A funny find is that the taildragger in the above example takes off at about 20-30 km/h less, with the exact same flight model setup! Angle of attack does exist in Arma.

 

 

But I really did not think of what you suggest. That sounds like a really good idea. Will let you know.

 

Share this post


Link to post
Share on other sites

not sure i see the solution to rear wheel steering yet - great work btw - this really needs sorting out

Share this post


Link to post
Share on other sites

Looks pretty good so far. Tried Reyhards suggestion. Having slight issues adjusting CoG as it has extreme effects on making the plane move on its own.

looks quite promising though. Thank you!

Share this post


Link to post
Share on other sites

Glad to hear that ;) I've used that trick on An-2 and it works decently so far. You can try it in RHS (or any other plane - all of them should have PhysX by now) and see how it works. I've also added engineMOI parameter to adjust how fast is plane able to accelerate on runway.

One thing that is very important is to match maxDrop & compression values  with what is model.cfg, so land contact points are moving correctly. I can give those models a shot during a weekend and leave them here if you want to see how it could be done.

  • Like 2

Share this post


Link to post
Share on other sites

thank you. that is very helpful. I made an update with a link in the first post.

A lot is lacking. I accidentally used my ill configured base as start instead of Tetete3s update. I added some of the values by hand. The result is quite ok though, judging from it's behaviour.

 

I found out that landcontact is not necessary for a flat tricycle gear setup. (you can even remove the wheels in geo lod...not sure what putting them there is good for...)

It helps a plane like the tail dragger though. It prevents the rolling away! (does not seem to be related)

 

Yes please feel free to play around with it! I will add a link for anyone interested then in first post.

Share this post


Link to post
Share on other sites

Kind of on a side note, but I think it's very relevant: None of your visible geometry have to have the same names as your virtual components. I found this useful when aligning the dampers/pistons and their stabilizers, the little elbows that keep the gears from turning out of position. Instead of using a translation animation approach in the model.cfg animations, use a rotation for your damper part - just like you would for those elbows.

 

In the skeleton section make the damper a child of the lower stabilizer elbow.

"damper_guide" ,"", //To provide something to inherit in animations. Might be redundant. 
"gear_1"	,"",  // The front gear mast, which lowers for landing. Connected to the root
"gear_1_2"	,"gear_1", // The first arm of the elbow, connected to the gear mast
"gear_1_3"	,"gear_1_2", // The second gear arm, its axis is the elbow joint. Connected to first arm
"gear_1_damper"	,"gear_1_3", //* The piston that goes in to and out of the gear mast. Connected to 2nd gear arm
"gear_1_caster"	,"gear_1_damper", //* Don't forget the steering on the front wheel.Connected to the piston. 
"visWheel_1"	,"gear_1_caster", //* The visible wheel; turns at virtual wheel's rate. Connected last.

// Steering can go before or after the three parts of the elbow assembly, but never within.
//* These have corresponding virtual components, and for this set up should be named differently. 


If your elbow parts are the same length, rotate the first elbow and the damper the same radian value, and rotate the middle (where it actually looks like an elbow) the negative of twice that amount. The goal is to nullify the difference. The piston will stay in its sleeve, and the joints will all stay "connected."

			class damper_guide_move ///again: this class might be redundant...
			{
				type="translation";
				source="Damper_1_source";
				selection="damper_guide";
				axis="basic_damper_destruct_axis"; //This is in a lot of vanilla vehicles
//				sourceAddress = clamp;// (default) ///these are in here, but why?
				minValue = 0.0;
				maxValue = 1;
				offset0 = 1;
				offset1 = -1;
				animPeriod = 0.0;
				initPhase = 0.0;
//				memory = true;//(default assumed)
			};
			
			class gear_1_stabil_1_rot: damper_guide_move ///...but we'll inherit from it anyway
			{
				type="rotation";
				source="Damper_1_source";///if inheriting from damper_guide_move, this might be redundant
				selection="gear_1_2";
				axis="gear_1_2_axis";
				minValue = 0.0; 		///redundant
				maxValue = 1;			///redundant
				angle0 = -0.0;
				angle1 = 1.0472;//rad 60; //Use also for Damper rotation when both elbow sections are equidistant. 
				animPeriod = 0.0;		///redundant
				initPhase = 0.0;		///redundant
//				memory = true;//(default assumed)
			};
			class gear_1_stabil_2_rot: gear_1_stabil_1_rot
			{
				selection="gear_1_3";
				axis="gear_1_3_axis";
				angle0 = 0.0;
				angle1 = -2.0944;//rad -120.0; //nullifies sum of angle changes for gear_1_2 and gear_1_damper
			};
			
			class gear_1_damper_rot: gear_1_stabil_1_rot //inherits most of gear_1_2 values
			{
				selection="gear_1_damper";
				axis="gear_1_damper_axis";
			};

It can still be done for landing gear stabilizers of different lengths, but it would require a little more math, knowing the lengths of the arms in the elbow, and thus knowing the initial and terminal angles. I believe the elbow angle should still be the negative of the sum of the other two angles. An example of this setup is in the A-10's Main Gear (rear). Despite the focus, the difference is clear here, too. What I did for my A-10 was I measured the angles in 3DSMax, at both extremes. I don't have actual specs on A-10 suspension stabilizer lengths, so I could fudge with it a little. Just make sure your angles add up to 180 in each position/phase, since it's technically a triangle. EDIT: After going back to look at my model rig, I remembered that the A-10 main suspension wasn't as simple as this. I had to set up an IK Arm Solver in 3DSMax, keyframe the damper going between its motion range extremes, and jot down the angles at both extremes. The problem arises with this one because the 2 axes of the elbow closest to the piston are not parallel to the piston. There probably is a mathematical way to sort it out, but it in this case it was easier for me to have the beginning and ending angles and then plug that data into the model.cfg. If you are having problems with a piece rotating the wrong way, then you can either A) try using the negative of whatever value you used, or keep the value but instead B) scale the axis to -1,1,1 in Object Builder MemLOD (if it's not on the X coordinate then  use -1,-1,-1).

I think I covered everything about this. The result should be a nice, connected, fluid motion on the suspension parts when dampering, whereas before it was misaligned or broken.

Share this post


Link to post
Share on other sites

Forgive me for the double post, but I wanted to distinguish these next thoughts from the previous one. Since that post was a little in-depth, I didn't want this to get glossed over.

1.) Let me impart some wisdom to you guys about case sensitivity. It was a major source of my problems which led to a bunch of unnecessary anguish and repeated trials. Even if you read an official statement somewhere claiming that a parameter is case insensitive, forget that noise and just stick to your principles. Always practice case sensitivity, and hunt down and assimilate rogue letters if you are pasting from someone else's script. For me, it was a 'w' where I needed a 'W' for Wheel_*. The W is sneaky for me, especially in the Courier-esque fonts that Notepad++ uses.

2.) This one is really more of a question, but first a brief outline of my trial and error process:

  • Start up A3, play-test, observe, shut down A3
  • Make changes to config.cpp, model.cfg, and/or [model].p3d
  • Compile in Addon Builder. If it fails without good explanation, compile in pboProject (free version) for better output. On successful compile in AB (needed for including .rtm files), move on.
  • Repeat process

It can be quite time-consuming, especially with all the load times. I feel like there should be a way to make changes in Eden to save time, and then use that data to make the changes permanent in the config (where most of the alterations take place). This would save time from all the load screens, not to mention how it would cut down on all the anguish over waiting, anticipating, and lingering disappointment accumulated between changes and results. I ran across this page: https://community.bistudio.com/wiki/Mission_Editor:_Debug_Console_(Arma_3). It seems to cover more about configuring for servers, but it does say changes can be made on the fly. It just doesn't specify what kind of changes - changes to an .sqf, .cpp, .cfg, or what? It seems to also be a fairly new feature, in the scheme of things. My question is: Has anyone used Eden's Debug Console to do fast edits to their config.cpp settings? Can it be done?

 

Edit: This question, though maybe not presented in the most efficient thread, is super important to anyone who has spent time tweaking airplane suspension. Thus, this is a good place for it. 

Edited by scotg

Share this post


Link to post
Share on other sites
6 hours ago, scotg said:

...

It can be quite time-consuming, especially with all the load times. I feel like there should be a way to make changes in Eden to save time, and then use that data to make the changes permanent in the config (where most of the alterations take place). This would save time from all the load screens, not to mention how it would cut down on all the anguish over waiting, anticipating, and lingering disappointment accumulated between changes and results. I ran across this page: https://community.bistudio.com/wiki/Mission_Editor:_Debug_Console_(Arma_3). It seems to cover more about configuring for servers, but it does say changes can be made on the fly. It just doesn't specify what kind of changes - changes to an .sqf, .cpp, .cfg, or what? It seems to also be a fairly new feature, in the scheme of things. My question is: Has anyone used Eden's Debug Console to do fast edits to their config.cpp settings? Can it be done?

 

 

The debug console can only use SQF to manipulate the game engine. What you're probably looking for is the diagnostic exe of the dev build: https://community.bistudio.com/wiki/Arma_3_Diagnostics_Exe

 

Of interest there is this command: https://community.bistudio.com/wiki/diag_mergeConfigFile

 

With help of diag_mergeConfigFile you can change and add values to an existing config from within the debug console. I have to admit that I rarely use it, but for all those detailed small adjustments on class wheel config values, it should be a real timer saver.

 

As converting your default game to dev build is not such a great idea, BI has published the game updater, that takes care of that for you: https://community.bistudio.com/wiki/Game_Updater

 

On your point 1), just discovered this case sensitivity when the wheels of a plane would refuse to roll in the game, with the wheels rolling perfectly in buldozer...

 

  • Thanks 1

Share this post


Link to post
Share on other sites

Just a note about devbranch - you can switch to devbranch, copy paste diag.exe somewhere, switch back to stable and use diag.exe along stable. As TeTeT mentioned, diag_mergeConfigFile should be your friend - you can merge config and i.e. respawn vehicle again to apply changes.

  • Thanks 2

Share this post


Link to post
Share on other sites

So now I'm trying to figure out how to use the diag_mergeConfigFile command. I did what @reyhard and @TeTeT said, esp. reyhard's advice about keeping a copy of the .exe before switching back to stable.

Based on the example here: https://community.bistudio.com/wiki/diag_mergeConfigFile, I replaced it with my plane's path:  diag_mergeConfigFile ["P:\gijoe\vhx_plane\cobra_firebat\config.cpp"]. Got an error message saying it was missing a ';' so I tried placing it at the end. Nothing. For unknown reason I tried placing it before the path. I would never have thought of that except for that one moment of randomness, but it did something. Started up scenario, and it didn't let me enter the plane. Weird. I didn't delete plane entry point; all I did was bump the spring strength up to to a ridiculous number, for the sake of testing the config merge. I'm guessing the missing ';' wasn't supposed to be where I put it, after all. For grits and shins, I shut down game and compiled with pboProject, but noisy output just gave me missing sound file warnings (i.e.: for sure no missing ';' in the config).

 

I think you guys are right about this saving me loads of time; I just need to figure it out first.

Share this post


Link to post
Share on other sites
15 hours ago, scotg said:

message saying it was missing a ';' so I tried placing it at the end. Nothing.

Semicolon at last line of script isn't needed. Look closely at where exactly the error messages hows it's missing.
Missing semicolon at:
diag_mergeConfigFile |#|["P:\gijoe\vhx_plane\cobra_firebat\config.cpp"]
Means that the script command doesn't exist.

Share this post


Link to post
Share on other sites
2 hours ago, Dedmen said:

Means that the script command doesn't exist.

Thanks, @Dedmen. So the |#| indicates where the semicolon goes? I did try putting it ahead of the file path, too. It did something. It didn't do anything right, but it did something.

Share this post


Link to post
Share on other sites

You're right. That must have been reyhard's shorthand for arma3diag_x64.exe. Diag.exe doesn't exist even in dev branch mode.

Share this post


Link to post
Share on other sites
On 1/9/2018 at 11:34 AM, DSabre said:

The TacOps update of November 2017 forces us to use wheel.cfg

This is confusing. @DSabre mentions wheel.cfg a few times throughout his first post, but I am not sure what wheel.cfg is. He has also mentioned it in other threads, if I recall correctly, and he has brought it up with me in private messages. Unless someone else knows what he's referring to, we may not be sure what this is until if/when he gets back into working on airplaneX stuff. It is taxing on one's sanity, especially if they feel like they're doing it alone. Getting this even just close to being right is a thing that really separates the big-team mod developers from individual, talented mod'ers.

Share this post


Link to post
Share on other sites

sorry that was me being up late.

what I meant is class wheels : )

 

sorry about that - and thanks for putting your thoughts and observations here

Share this post


Link to post
Share on other sites

Ah, ok. It's class wheels in the config.cpp. That I do have. Things are still a bit wonky, so if I start to think I'm missing something it'll haunt me.

Share this post


Link to post
Share on other sites

I fixed some mistakes in my first post.


Another thing I just observed while updating my navy planes to work better with the carriers:

 

I copy/pasted my taildragger A6M Zero setup into my SBD Dauntless, TBF Avenger, Aichi Val and the Nakajima Kate, adjusting wheel positions to match the model.

That worked mostly well. When I put it in the Avenger and Val they would just roll all over the place though.

I played around for hours moving CoG here and there. Nothing helped. Again the frustration....of fixing things that had worked well before...

 

This is what finally helped:

simply moving the landcontact of the tail wheel up into the wheel position. I believe that acts as some kind of brake.

 

What caused the rolling off compared to the other aircraft was the position of Tail wheel in relation to the front gear.

Not sure what to make of it. But then taildraggers are not properly supported anyway and all of this is some kind of hack : )

 

 

 

3F7CAB70698C6CD49281F0BD0BAF8C045AE9FE56

 

 

 

physxtbf.png

 

Overlay of memory, geo lod and landcontact lod items in an edit lod (I use this to align and then copy/paste into the respective layer):

The physx mem wheel points are where you would expect them to be.

The LC points are a bit to the front of the wheel and a bit above the actual contact.

 

Note that I use wheel_l,r and c (left, center, right) instead of wheel_1,2 and 3.

I never know which is the left or center wheel with the stupid numbers.

 

Both geo lod wheels and memory lod wheels are smaller than the actual resolution lod wheel. That's my way of preventing the model from visually floating over the ground.

You could of course set up all dampers and animations and things precisly - but with over 100 aircraft I just do not have the time for this nonsense. Considering they all used to work just fine before last winter.

 

 

 

Incase someone is interested, some more details:
 

 

Spoiler

https://www.dropbox.com/s/7uezqgnpsqqjqvf/avenger physx setup.p3d?dl=0

(link to the geo lod, mem lod and lc lod setup)

 

 

CoG is between the 2 wing mass points. total mass my standard of 1500 units.

   

 

physxgeo.png

 

 

 

 

class Wheels {
            
            class Wheel_c {
        
                steering = 0; //thanks reyhard!, tail dragger hack
                
                boneName = "wheel_c";  // linked for visual display of droop and compression?
                center = "wheel_c_center";
                boundary = "wheel_c_rim";
                suspForceAppPointOffset = "wheel_c_center";
                tireForceAppPointOffset = "wheel_c_center";
                                    
                width = 0.2;
                mass = 10;
                MOI = 1;
                
                dampingRate = 0.1; // 0.1
                dampingRateDamaged = 10;
                dampingRateDestroyed = 1000;
                maxBrakeTorque = 100;   // put a million in here... no effect at all
                maxHandBrakeTorque = 100; // same, no effect..
                suspTravelDirection[] = {0, -1, 0};  // this gentlemen, is the actual physx dampers movement direction. in relation to your center and boundary wheel points I believe.

                
                maxCompression = 0.01; // how far the virtual damper can compress
                maxDroop = 0.01;        // how far the virtual damper moves down without load, I think. Some say we have to make the axis in mem lod the exact length? This does not make sense to me. that is like a conflicting definition
                longitudinalStiffnessPerUnitGravity = 1000;  // all these values are from trial and error. the formulas did not do it for me
                latStiffX = 2;
                latStiffY = 3;
                frictionVsSlipGraph[] = {{0, 1}, {0.5, 1}, {1, 1}};
            
                //sprungMass = 500;  // all these values are from trial and error. the formulas did not do it for me
                //    springStrength = 5000;  // all these values are from trial and error. the formulas did not do it for me
                //springDamperRate = 6000; // all these values are from trial and error. the formulas did not do it for me
                
                sprungMass = 700; //all these values are from trial and error. the formulas did not do it for me
                springStrength = 4000; //all these values are from trial and error. the formulas did not do it for me
                springDamperRate = 5000; //all these values are from trial and error. the formulas did not do it for me
                
            };
            
            class Wheel_l : Wheel_c {
                MOI = 0.1;  // 0.1
                steering = 1; //thanks reyhard! tail dragger hack
                maxCompression = 0.01; // 0.01
                maxDroop = 0.01;        // 0.01
                maxHandBrakeTorque = 0;
                maxBrakeTorque = 0;
                boneName = "wheel_l";
                side = "left";
                center = "wheel_l_center";
                boundary = "wheel_l_rim";
                suspForceAppPointOffset = "wheel_l_center";
                tireForceAppPointOffset = "Wheel_l_center";
            };
            
            class Wheel_r : Wheel_l {
                
                boneName = "wheel_r";
                side = "right";
                center = "wheel_r_center";
                boundary = "wheel_r_rim";
                suspForceAppPointOffset = "wheel_r_center";
                tireForceAppPointOffset = "wheel_r_center";
            };
        };
   

 

  • Like 1
  • Thanks 1

Share this post


Link to post
Share on other sites
1 hour ago, DSabre said:

Note that I use wheel_l,r and c (left, center, right) instead of wheel_1,2 and 3.

I never know which is the left or center wheel with the stupid numbers.

I always assumed, for planes anyway, that the order is front to back, left to right. So in a tricycle plane wheel_1 would be the front wheel, wheel_2 would be the left main wheel, and then wheel_3 is the right main. In taildragger layout I would think wheel_1 is left main, wheel_2 right main, and wheel_3 is the rear wheel. That's my assumption, but since I haven't tried making a taildragger it has little merit.

Any idea how many wheels total per plane are supported in A3?
Tanks are something like 10 ground wheels per side, and I think cars are 8 in total... or at least I haven't seen anything more than 8.

Share this post


Link to post
Share on other sites

I would suggest to model 3 wheels max with physX and animate any surplus wheels only in res lod. That should save a lot of work

So I have no idea how many would be in theory possible..

... maybe, just maybe I will try on the An-22, just to see what happens : )

 

Edited by DSabre
never say never

Share this post


Link to post
Share on other sites
On 10/4/2018 at 1:09 PM, DSabre said:

I would suggest to model 3 wheels max with physX and animate any surplus wheels only in res lod. That should save a lot of work

So I have no idea how many would be in theory possible..

... maybe, just maybe I will try on the An-22, just to see what happens : )

 

I was re-reading this thread to see if there were any further developments, and because I reached out to you in a PM. That aside for this thread, I haven't been around much since using the Discord server more. I'm curious if you ever did the An-22, or any other plane with multi-wheel gear. Some larger aircraft have 8 wheels per main gear. The ones on the back end would teeter totter the entire wheel assembly before the gear suspension compresses on landing. Simply put, the way a (tricycle) plane lands on its main gear first and then sets down on its front gear, the row of wheels on one gear would behave the same way. On bigger planes, not showing this behavior would be ugly.

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now

×